Introduction

Injuries in sport are costly [8] and preventing injuries can benefit sporting teams by improving both player availability for games and team performance in competitions [22]. While there are many possible modifiable and non-modifiable risk factors for injury, an association between pre-season measures of size and function of trunk muscles and future lower limb injuries has been previously identified [11]. The relationship between size of the multifidus (MF), psoas major, quadratus lumborum (QL) and the abdominal muscles and injuries was first investigated in players from the Australian Football League (AFL) [12]. Results showed that players with more severe injuries (reflected by more games missed) had significantly smaller MF muscles compared with players without injuries.

While the actual mechanism underpinning the relationship between the size and function of trunk muscles and injuries remains unknown, several theoretical explanations have been put forward. It has been proposed that suboptimal force transfer could contribute to injuries in athletes with deficient muscles in the lumbo-pelvic region. In support of this theory, prospective laboratory studies that investigated neuromuscular control of the trunk have shown that deficits in this area predicted lower limb injuries [44, 45]. The proposed rationale for the finding was that decreased neuromuscular control of the trunk, coupled with high ground reaction forces directed toward the body’s centre of mass, compromised the dynamic stability of the knee joint and increased knee injury risk. Researchers investigating upper limb injuries and altered neuromuscular control of trunk muscles have also proposed that well-functioning and co-ordinated muscle systems are required to control limb and trunk movement and protect players from injuries [15, 40].

Recently, a study of AFL players confirmed the relationship between the size of the MF and QL muscles and an increased risk of lower limb injury in players from the AFL [9]. The models positively predicted 75% of players that sustained a lower limb injury during the playing season (sensitivity 80%, specificity 85%). Clinical cut-off values of < 8.5 cm2 of the MF muscles at the L5 vertebral level, and > 8.2 cm2 of the QL muscle have been previously reported [20]. These clinical cut-off values, derived from AFL players, have since been applied to players from the AFL and the National Rugby League (NRL) [7]. In support of previous results, players from the AFL with larger QL muscles and a lower ratio of the size of the MF to the QL muscle (AUC 0.7–0.8) had increased odds of sustaining a lower limb injury during the playing season [7]. In contrast to previous studies, relationships were not found between the size of the MF muscles and lower limb injuries in AFL or NRL players [7]. However, given that the clinical cut-off values for MF muscle size were derived from AFL players, the applicability of these measures to athletes from other sports (such as rugby league) represents a current gap in the literature.

Whilst the majority of studies of trunk muscle size and function have investigated the relationship between trunk muscles and lower limb injuries, relationships between these muscles and upper limb and head/neck and concussion injuries have also been identified. A study of rugby union, rugby league and Australian football players showed that trunk muscle size and function predicted season head/neck injuries and concussion injuries [14]. Players with a history of head and neck injuries were found to have decreased size of the MF muscles and larger QL muscles [14] and altered ability to contract the lumbar MF muscles also predicted head and neck injuries [16]. Interestingly, rugby league, rugby union and AFL players who sustained a season head/neck injury contracted their MF muscles more than players who did not sustain an injury [14], which has also been observed in other prospective studies of concussion [13, 26]. For the anterolateral abdominal muscles, a pattern of increased contraction was observed for the transversus abdominis muscle [14], and increased thickness of the internal oblique muscle was observed post-concussion [13].

Possible rationales for a relationship between the ability to contract the trunk muscles and head and neck injuries in contact sports may be related to factors including the mechanism of injury, the role of the trunk muscles and the relationship between trunk and neck muscles [16]. Many concussion injuries occur during tackling [4]. Spinal muscles such as the MF play an important sensory as well as a biomechanical role in controlling segmental motion of the lumbar spine and the lordosis [3]. The position of the trunk when tackling is important, with correct foot placement and optimal positioning of the trunk identified as important elements of a safe tackling technique [39]. Correct activation of the MF muscles would therefore be required to enable positioning of the lumbar spine in an optimal position to allow adequate distribution of the forces sustained on impact in contact sports.

The researchers who conducted the current investigation were approached by a professional rugby league team to determine if a relationship between trunk muscle size and function was related to injuries in their club. If a relationship existed, the medical team were interested in determining clinical cut-off values appropriate for the rugby league players at their club. The aims of this study were to investigate if a) trunk muscle size and function were associated with an increased risk of playing season injuries (reflected by the number of games missed due to injury) and if b) trunk muscle size and function were associated with an increased risk of concussion injuries in a team of professional rugby league players. The study also aimed to determine clinical cut-off values related to size and function of trunk muscles.

Methods

Participants

A prospective observational study of players from a professional rugby league club in Queensland, Australia was conducted. Players aged 18 years and over, who were part of the professional squad, were eligible for participation in the study. Participants were assessed during the 2015 pre-season, and injury data from the following three seasons were recorded. Written informed consent was obtained from all individual participants included in the study. This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Human Research Ethics Committee of Griffith University (2017/864).

Trunk Muscle Imaging

Ultrasound imaging of the trunk muscles was performed in the rugby league pre-season. The protocols have been published in full elsewhere [18, 20]. Ultrasound imaging was conducted using LOGIQ e apparatus with a 5-MHz curvilinear transducer (GE Healthcare, Wuxi, China). The MF muscles were imaged bilaterally from the L2 to L5 vertebral levels, (Fig. 1a) with the player positioned in prone lying [18]. The QL muscles were imaged bilaterally in line with the L3–4 vertebral interspace [20]. The MF, transversus abdominis and internal oblique muscles were imaged both at rest and on contraction (Figs. 1c and d). Ultrasound images of the muscles of the antero-lateral abdominal wall were obtained with the player positioned in supine lying. A transverse image was obtained along a line midway between the inferior angle of the rib cage and the iliac crest [17]. The transducer was aligned perpendicular to the fascia covering the anterolateral abdominal muscles. For assessment of contraction of the abdominal muscles, participants were asked to draw in the lower abdomen without moving the spine [17]. To assess contraction of the MF muscle, imaging was performed in parasagittal section, allowing visualisation of the L2/3, L3/4, L4/5 and L5/S1 zygapophyseal joints, muscle bulk and thoracolumbar fascia. Participants were instructed how to perform a voluntary isometric contraction to familiarise them with the contraction required for testing. This involved the examiner placing their fingers on the player’s back and giving them one practice contraction. The player was instructed to “relax the trunk muscles, then take a relaxed breath in, and out, hold the breath out and then to try to contract or swell the multifidus muscle into the examiner’s fingers.” The MF muscles were then imaged at rest and on contraction using this instruction from the L2 to the L5 vertebral levels on both sides.

Fig. 1
figure 1

a Bilateral transverse ultrasound image of the multifidus muscle, with the borders of the multifidus outlined b to demonstrate segmentation. ST subcutaneous tissue, SP acoustic shadow of spinous process, MF multifidus muscle, L5 vertebral level, L acoustic shadow of the lamina. c Parasagittal ultrasound images of the multifidus muscle, using a split screen, with muscle thickness measured from the top of the zygapophyseal joint to the thoracolumbar fascia in relaxed (left of image) and contracted (right of image) conditions. L4/L5 and L5/S1 indicate zygapophyseal joints. MF multifidus muscles, ST subcutaneous tissue. d Transverse ultrasound image of the muscles of the anterolateral abdominal wall, in relaxed (left of image) and contracted (right of image) conditions, using a split screen. Lines indicate thickness measures of the transversus abdominis and internal oblique muscles in the relaxed and contracted states. EO external oblique muscle, IO internal oblique muscle, TrA transversus abdominis muscle, ST subcutaneous tissue

Full size image

Ultrasound images were stored and measured offline using OsiriX medical imaging software (Geneva, Switzerland). Two experienced researchers with over twelve years of experience in measuring the muscles of interest and demonstrated intra-rater reliability conducted the measurements of QL size (ICC = 0.99) [20], MF muscle size (ICC mean for L2–L5 = 0.94) [18], MF muscle thickness (ICC = 0.88–0.95, relaxed and contracted) [42], and abdominal muscle thickness (transversus abdominis ICC = 0.62–0.98; internal oblique ICC = 0.69–0.99, relaxed and contracted) [17]. Inter-rater reliability for measurements of the MF, transversus abdominis and internal oblique muscles was high (ICCs ranging from 0.88 to 0.99) [10]. To determine the percentage muscle contraction, the difference between the contracted and relaxed measures was expressed as a percentage of the resting value. The researcher who conducted the measurements was blinded to player injuries, and measurements were conducted and provided to the club before injury reporting was conducted.

Season Injury Data Collection

Injury data were collected by the medical team for three seasons (2015–2017) following the collection of muscle imaging data. The data were provided to the research team at the end of the study period. An injury was defined as a physical condition related to playing or training for football resulting in a musculoskeletal injury to the body that prevented a player from playing in subsequent games. Concussion injuries were diagnosed and recorded by medical staff.

Statistical Analysis

SPSS version 28.0 (IBM) was used for analyses. Data were examined for outliers and normal distribution using summary statistics, histograms, normality plots and the Kolmogorov–Smirnov test. The measurements tested as risk factors included both categorical and continuous variables (age, height, mass and trunk muscle size and contraction measures). All factors were assessed against injuries in the 2015–2017 playing seasons (reflected by the number of games missed) and concussion injuries. Because the majority of players missed at least one game due to injury in the three-year time period studied, the ‘injury’ group was categorised as ‘those who missed more than five games due to injury over the 3-year period’. The cut-off value for severity was based on a criterion of achieving close to equally sized injury groups. Participants have been grouped this way in previous AFL studies, where greater than four games missed was used to categorise players [12]. Univariate associations between categorical variables were examined using chi-square tests. Continuous variables of age, height and mass were analyzed using one-way ANOVAs, with the between group factor being either of the injury variables. One-factor repeated-measures ANOVA was conducted to compare the continuous variables (trunk muscle size and contraction measures) between groups (0–5 games missed/ > 5 games missed and concussion/no concussion). If a significant relationship existed, receiver operating characteristic (ROC) curves were used to determine an optimal cut point for each continuous variable in predicting the injury outcome variables. The optimal cut point was obtained as the point where the true-positive rate (sensitivity) was maximized and the false-positive rate (1 – specificity) was minimized, that is, the point closest to the top of the y-axis. The cut points were then used to convert the continuous variables to a binary form. The sensitivity and specificity of each measure and the unadjusted odds ratio (OR) for predicting injuries and concussions in the playing season were estimated from cross-tabulations and chi-square tests. Using a forward sequential method to conserve statistical power and to investigate possible collinearity between variables at each step [1], significant binary variables were entered into logistic regression models in order of their univariate effect (P value) for predicting injuries (0–5 games missed/ > 5 games missed) and concussion injuries in the playing season. Due to the small sample size, only two predictors could be included in each model. To minimise the possibility of Type II errors, variables were retained in the model if P < 0.1 [23]. Because this was an exploratory study, no adjustments for multiple comparisons were made to avoid Type II errors [2]. Variables that were not significant multivariate predictors or that were collinear with other variables were dropped from the models.

Results

Twenty-eight male rugby league players were included in the study. Mean (SD) participant age (yrs), height (cm), and mass (kg) were: 25.1 (3.6) years, 185 (4.7) cm, and 101 (8.6) kg.

Number of Season Games Missed and Types of Injuries

In the 2015–2017 seasons, 17 players missed five or less games, with a mean (SD) of 1.3 (1.7) and 11 players missed more than five games due to injury, mean (SD) = 20 (11.2). In the group that missed 0–5 games, the types of injuries recorded included muscle strains/ tears (hamstrings, gluteus medius, rectus femoris, adductor and soleus muscle); joint and ligament injuries (acromioclavicular joint, shoulder, knee, ankle) and concussion injuries. The types of injuries reported in the greater than 5 games missed included muscle strains/ tears (hamstrings, rectus femoris, soleus, pectoralis major); joint and ligament injuries (cervical spine, shoulder, sternoclavicular, hip, knee); tendon injuries (Achilles, adductor); fractures (tibial, hand) and concussion injuries. Of the 28 players, a total of 12 (43%) players sustained a concussion in the 2015–2017 playing seasons. There were no significant differences between the injury groups and concussion groups for age, height or mass (all P > 0.05).

Games Missed due to Season Injuries and Muscle Size and Function

Games missed due to season injuries were related to trunk muscle size (Table 1) and trunk muscle contraction (Table 2). Players with > 5 games missed had smaller MF muscles at the L5 vertebral level (L5L, P = 0.022, Effect size: ES = 0.92) and smaller QL muscles (QLR, P = 0.048, ES = 0.86; QLL, P = 0.094, ES = 0.68). With respect to muscle contraction, players who missed > 5 games due to injury contracted their MF muscles more at the L2/3 level (MFR, P = 0.028, ES = 0.44; MFL, P = 0.08, ES = 0.65;) and more asymmetrically at the L4/5 vertebral level (P = 0.094, ES = 0.68), with greater contraction on the right side, compared with players who missed 0–5 games (Table 2). Measures with a significant area under the curve (AUC) are shown in Table 3. The measures with the highest sensitivity and specificity for in-season injuries (> 5 games missed) were CSA of the MF and QL muscles. The unadjusted odds ratios and adjusted odd ratios (calculated using logistic regression) for the risk factors associated with playing season injuries are shown in Table 4. The adjusted OR values in Table 4 indicate that if a player has a small MF muscle (< 9.98 cm2 at L5; P = 0.032) or a small QL muscle (< 10.8 cm2; P = 0.045), their odds of an injury resulting in games missed during the season would be significantly increased. Of the 28 players assessed, eight did not possess either risk factor, eight possessed both risk factors and 12 possessed one risk factor. Only two players with both risk factors did not have an injury (25%), compared with six of the eight (75%) players with two risk factors who did have injuries (P = 0.03 with an OR of 6). This suggests that the odds of a player with two of the risk factors having a season injury resulting in games missed is six times higher than that of a player with one or no risk factors. Using the criterion of two risk variables results in a sensitivity of 75% and a specificity of 75% for predicting season injuries (> 5 games missed).

Table 1 Trunk muscle size for players who missed 0–5 or > 5 games due to injury in the 2015–2017 playing seasons
Full size table
Table 2 Measurements of trunk muscle size and function for players who missed 0–5 or > 5 games due to injury in the playing season
Full size table
Table 3 ROC curves optimal cut values for variables associated with either > 5 games missed or experiencing concussion injury during the season
Full size table
Table 4 Unadjusted and adjusted odds ratios for multifidus and quadratus lumborum muscle size (as risk factors for injuries resulting in > 5 games missed in the playing seasons) and multifidus and abdominal muscle function (as risk factors for concussion injury)
Full size table

Concussion Injuries and Muscle Size and Function

Players who experienced concussion injuries did not demonstrate differences in trunk muscle muscle size (Table 5) but had greater contraction of the right MF muscles at the L4 (L4R, P = 0.028, ES = 0.7), and L5 vertebral level (L5R, P = 0.028, ES = 0.85) (Table 6). For the anterolateral abdominal muscles, a similar pattern of increased contraction was seen on the right side for the transversus abdominis muscle (P = 0.072, ES = 0.66) and on the left side for the internal oblique muscle (% change P = 0.057, Table 6). The measures with the highest sensitivity and specificity for concussion injuries were contraction of the MF and abdominal muscles (Table 3). The adjusted ORs in Table 4 indicate that if a player had a larger contraction of the MF muscle (> 7.2% at L4; P = 0.028) or a larger contraction of transversus abdominis muscle (> 49.9%; P = 0.08), their odds of a concussion injury during the season would be significantly increased. Of the 28 players assessed, seven did not possess either risk factor, 12 possessed both risk factors and nine had 1 risk factor. Eighty-three percent of players with two risk factors experienced a concussion injury. Players without risk factors or with one risk factor for concussion had lower odds of sustaining a concussion (P = 0.09) with an OR of 3.1. Using the criterion of no risk variables for concussion results in a sensitivity of 52.4% and a specificity of 85.7% for predicting concussion injuries.

Table 5 Trunk muscle size for players who did and did not sustain a concussion injury in the 2015–2017 playing seasons
Full size table
Table 6 Measurements of trunk muscle size and function for players who did and did not sustain a concussion injury in the playing season
Full size table

Discussion

The main result of the current investigation indicated that smaller size of the MF and QL muscles was associated with injuries resulting in games missed in the playing season in a team of professional rugby league players. With respect to concussion injuries, results indicated that increased contraction of the MF and abdominal muscles was associated with increased odds of sustaining a sports related concussion.

Games Missed due to Season Injuries and Muscle Size and Function

The results of the current investigation are consistent with studies of other sports that have also demonstrated a relationship between the multifidus muscles and injuries. In soccer players, Nandlall et al. found that greater atrophy of the MF muscles over the playing season was associated with season lower limb injuries [34]. A study of university rugby union players showed that those with a lower limb injury in the previous 12 months had greater between side asymmetry of the cross-sectional area (CSA) of the MF muscles [27]. Another study of rugby union players identified that decreased ability to contract the lumbar MF was associated with lower limb injuries [37]. A recent study reported that rugby union players with upper limb injuries had smaller MF and more asymmetrical MF muscles at the L5 vertebral level [28]. A study of volleyball players also demonstrated a relationship between previous upper limb injuries and an altered ability to contract the MF muscle [15]. Previous studies of AFL players have also shown that smaller MF muscle size was associated with injuries in the AFL playing season [11, 12, 19]. However, the clinical cut-off of 9.98 cm2 determined for rugby league players in the current investigation was higher than that previously determined for AFL players (8.5 cm2) [20]. This finding may help to explain the difference in results between the current investigation and those of Hajek et al. [7], who did not find a relationship between MF CSA and injuries. Because NRL players have larger muscles than AFL players, the cut-off derived from AFL players may not be appropriate to adopt for use in players from the NRL [7].

The results for the size of the QL muscle in relation to season injuries in the current investigation showed that having smaller QL muscles increased the risk of injuries. This result supports one previous study [19] but differs from other previous studies that have shown that increased size of the QL muscle predicted injuries in Rugby League [7] and AFL players [9, 20]. The possibility of a smaller QL muscle being associated with injury could be related to the muscle not being able to adequately perform its proposed role of controlling spinal buckling [30]. However, the actual role of the QL muscle remains controversial [35]. It has been proposed that the QL muscle plays an important role in stabilization of the lumbar spine [31], which has led to interventions aimed at increasing recruitment of the QL muscle [29, 33]. In contrast, many clinical treatments focus on decreasing the activation of the QL muscle [6, 38]. It is possible that the differing results regarding the relationship between the QL muscle and injury could represent a spectrum and adaptation to different sporting activities. In contact sports, contraction of the QL could be beneficial when tackling, as this could result in stiffening of the spine, suggesting that a larger QL could be protective. However, it is important that the human body is not ‘rigid’ during contact to allow absorption of forces. Increased size and activation of the QL muscle could possibly represent a disadvantage during running, as the spine needs to be able to laterally flex during this activity, and dominance of the QL bilaterally could result in lack of lateral flexion of the lumbar spine, rigidity and an inability to absorb shock and distribute loads effectively [20]. In line with this finding, a review paper proposed that when back muscle fatigue occurs, muscles such as QL increase activity to compensate, resulting in a cascade of events which could potentially result in lower limb injuries [5].

Concussion Injuries and Muscle Size and Function

We found that players who sustained a concussion in the playing seasons had greater contraction of the MF muscles at the L4 and L5 vertebral levels and increased contraction of the transversus abdominis and internal oblique muscles, which supports the results of two previous studies of the MF muscles in rugby union, rugby league and Australian football players [13, 26]. Regarding the anterolateral abdominal muscles, a pattern of increased contraction was also observed in the transversus abdominis muscles in those who went on to sustain a concussion injury [14].

Considering the results of studies of football players pre- and post-concussion may help explain these findings. Increased size and contraction of trunk muscles has been observed 3–5 days post-concussion [14, 26]. It was proposed that these findings may represent a strategy of splinting or over-holding possibly leading to increased trunk stiffness [14, 26]. A similar response has been observed in people with low back pain (LBP), where researchers have proposed that higher stiffness among individuals with LBP, even when not experiencing pain, may result from higher baseline electromyographic levels in the trunk musculature [25, 36, 41, 43]. While these compensatory responses may represent appropriate strategies in the acute situation, they may not be optimal long-term strategies, as they may interfere with normal movement [24] and increase spinal loads [32]. It is possible that players with increased contraction of their trunk muscles may have had higher odds of experiencing a concussion if they were less able to distribute forces through the body during high impact activities such as tackling.

Clinical Implications

Based on the observation that neuromuscular control of trunk muscles may be associated with injuries across several sports, it has been proposed that exercise strategies targeting hypertrophy of trunk muscles, such as the MF, may be potentially beneficial for injury prevention. It has been confirmed that the size of the MF muscle is modifiable in response to exercise in AFL players [21]. Results showed that AFL players who trained their MF muscles as part of a staged, targeted exercise program in the playing season missed fewer games due to injury compared with players from a control group [21]. However, interventions targeting trunk muscles with an aim of decreasing concussion injuries have not yet been tested. Establishing a clinical cut-off for rugby league players for trunk muscle size would represent a step towards identifying the players (beneath the clinical cut-off value) who may potentially benefit the most from undertaking a motor control training program targeting the trunk muscles.

Limitations

The main limitation of the current investigation was that a single NRL club was assessed in the current investigation. The small sample size limited the number of potential risk factors that could be examined statistically. The selection of zero to five games or greater than five games missed as a grouping variable reflecting severity of injuries was arbitrary in this study. The wide confidence intervals of the odds ratios in the current exploratory investigation are likely to be related to the small sample size—and may indicate that while the factors identified can be considered as increasing the likelihood of injury, their precision may be limited. The cut point values determined in this study may not be transferable to other rugby league teams. Future studies with larger sample sizes and players from several NRL clubs would allow a more thorough investigation of a greater number of potential risk factors.

Conclusion

The results of studies conducted by various research teams in different sports have suggested that changes in trunk muscle size and function are related to sports injuries. This finding perhaps should not be surprising when the important role of the trunk in function is considered. However, muscle measurements are not perfect predictors, and for some sports, the results appear conflicting. This is perhaps because there are likely to be several other modifiable risk factors (apart from trunk muscle size and function) affecting athletes that influence whether injuries occur. It appears that sport-specific assessments may be required, and that clinical cut-off values are not transferable between different sports and teams. The concept of a spectrum of changes in muscles (such as the QL) adds complexity to planning optimal interventions in clinical trials or team environments where injury prevention is the aim. In the AFL, there is evidence that undertaking exercises targeting impaired muscle size and function was associated with decreased games missed due to injury. It is currently unknown whether addressing the identified impairments in trunk muscles in rugby league players will decrease injury incidence, as was observed in AFL players, as the morphology of the players and the requirements of the two sports are very different. Further investigation is therefore required.